Objective：To assess whether the first blood lipid test in outpatients is necessary to detect low-density lipoprotein cholesterol（LDLC） in order to achieve grading test of medical examination，reduce the labor of medical inspectors and reduce medical expenses. Methods：Total cholesterol（CHO），triglycerides（TG），high-density lipoprotein cholesterol（HDLC），and LDLC data were obtained from Laboratory Information System（LIS） based on outpatients in 4 years（2013-2017）. CHO，HDLC and LDLC were measured using TBA2000FR biochemical analyzer. nonHDLC was calculated with CHO minus HDLC. Correlation between CHO，nonHDLC and LDLC were analyzed using Spearman’s rank approach. Receiver operating characteristics（ROC） curve analysis was used to evaluate the predictive of CHO and nonHDLC for abnormal LDLC. Results：Both CHO（r=0.843）and nonHDLC（r=0.862）were significantly positively correlated with LDLC. Area under curve of CHO and nonHDLC for predicting abnormal LDLC（＞3.40 mmol/L and＞4.10 mmol/L）were 0.941 and 0.948，0.967 and 0.970，respectively（P＜0.001）. Optimal thresholds of prediction abnormal LDLC were 3.85 mmol/L and 4.33 mmol/L for nonHDLC. Based on these optimal thresholds，the sensitivity of nonHDLC to predict LDLC positive results was 91.1% and 97.6%，and the specificity was 85.4% and 87.3%，respectively. Less than 3.33% and 0.16% of tests with abnormal LDLC might be missed，but approximately 66.3% and 80.6% of the LDLC tests could be eliminated. Conclusion：We recommend that LDLC measurement is not necessary for the first test of blood lipids in outpatients. About 65% of LDLC tests would be reduced. Moreover，nonHDLC is a better index than CHO to predict abnormal LDLC.